BACKGROUND: Young children of pre-school age may find a minimal intervention (fluoride varnish application) difficult to tolerate. AIM: To determine the significant predictors for refusing a fluoride varnish application from child, parental and nurse behaviour factors. DESIGN: Data included videos from 238 children (52% female, aged 3-5 years) receiving a fluoride varnish application in a Scottish nursery school setting. The St Andrews Behavioural Interaction Scheme (SABICS) was used for video coding and retrieved child refusal status, initial anxious behaviour, and nurse behaviour. A parental survey collected parent’s dental anxiety [Modified Dental Anxiety Scale (MDAS)] and the child’s home behaviour [Strengths and Difficulties Questionnaire (SDQ)]. Child demographics, dental status, and previous varnish application experience were recorded. Multivariate binary logistic regression was applied to predict child refusal of the varnish application. RESULTS: The response rate was 79%. Twelve children refused. The significant predictors of varnish refusal included initial anxious child behaviour (beta = 5.14, P = 0.001), no previous varnish application (beta = -3.89, P = 0.04), and no nurse praise (beta = -1.06, P = 0.02). Information giving (P = 0.06) and reassurance (P = 0.08) were borderline significant. CONCLUSION: Initial anxiety behaviour, previous varnish experience, and not using praise by the nursing staff predicted fluoride varnish application refusal.